Emotion recognition method and device for audio data, equipment and medium

• Main Authors: RU GANGHUI, WANG JIANZONG, ZHANG XULONG
• Format: Patent
• Language: Chinese; English
• Published: 18.08.2023
• Subjects: ACOUSTICS; MUSICAL INSTRUMENTS; PHYSICS; SPEECH ANALYSIS OR SYNTHESIS; SPEECH OR AUDIO CODING OR DECODING; SPEECH OR VOICE PROCESSING; SPEECH RECOGNITION

The invention is suitable for the technical field of artificial intelligence, and particularly relates to an emotion recognition method and device for audio data, equipment and a medium. The method comprises the following steps: performing sound source separation on acquired audio data by using a sound mode extraction model to obtain separated audio of at least one sound mode, comparing the sound mode of the separated audio with sound modes in a first mapping table, determining an encoding model corresponding to the separated audio, and calling the corresponding encoding model to encode the separated audio. Connecting all coding results, inputting a connecting result into a decoding model, inputting a decoding result into a classification model, outputting an emotion category of audio data, performing different coding processing on separated audios by using different models through audio separation, and integrating the separated audios to output an emotion classification result. The method can achieve the pro